Low-latency, low-power audio algorithms — classical DSP and modern ML — running where the silicon is small and the milliwatt budget is smaller. Principal inventor on patents shipped in billions of earphones.
A browser MIDI arpeggiator for Korg Berlin's phase8 — the eight-voice acoustic synthesizer with steel resonators. Twenty-eight hand-authored generative compositions (Bach, Reich, Eno, Pärt, Cage, Satie…), each a different musical territory using one engine. Runs in Chrome over Web MIDI — no installer, no DAW.
Co-founded loudspeaker company. Engineered an all-carbon-fibre active tower with a novel waveguide aimed at the listening seat to suppress room coloration. Voiced in collaboration with composer / producer Amon Tobin.
Speak a wish at a Raspberry Pi → Whisper transcribes → GPT writes a fresh Arduino sketch → the lights reprogram themselves. Conversational firmware, end-to-end.
Programmable LED controller running across an ad-hoc Wi-Fi mesh of Pi + Arduino nodes. No router, no cloud — just peers.
An older demo of microphone-array beamforming — picking out a single voice in a noisy room without losing the room.
Long-form fiction. Audio engineers do other things too.